PRENATAL DIAGNOSIS, VOL. 14 455-458 (1994)

SCREENING MATERNAL SERUM ALPHA-FETOPROTEIN LEVELS AND HUMAN

PARVOVIRUS ANTIBODIES DAVID R. JOHNSON*$, RACHEL A. FISHER?, JENNY J. HELWICK?, DENNIS L. MURRAY?, MARIA J. PATTERSON?

AND FRANCES P. DOWNESS

*Division of Field Epidemiology, Centers for Disease Control, Atlanta, GA 30333. U. S. A.; ?Department of Pediatrics and Human Development, Michigan State University, East Lansing, MI 48824, U. S. A.; $Bureau of

Infectious Disease Control, Michigan Department of Public Health, Lansing, MI 48909. U.S. A.

Received May 1992 Revised September 1993 Accepted October I993

SUMMARY The association between gestational infection with human parvovirus (B 19) and fetal loss has increased interest in

this virus and demand for diagnostic testing. However, serological assays for B19 are not yet widely available. Maternal serum alpha-fetoprotein (MSAFP) testing is commonly used during the second trimester to screen for various fetal defects. We attempted to determine whether an elevated level of MSAFP would be an appropriate indication for B19-specific tests. Over a 26-month period, MSAFP tests were performed at Michigan State University for 21 392 women. Sera remaining after that testing were stored frozen. Of these, 22 case samples-from women with MSAFP levels greater than 3.0 multiples of the median (MOM) and pregnancies that ended in fetal loss-and 44 matched control samples-from women with MSAFP levels greater than 0.4 and less than 2.2 MOM and live births at term-were tested for B19 antibodies. None of the 66 samples was IgM positive, while 33 (50 per cent) were IgG positive. The presence of IgG was not significantly associated with case or control status (matched odds ratioz0.77, 95 per cent confidence interval 0-28-2.11). These findings are consistent with other studies indicating prior infection in approximately half of adults and suggest that elevated screening MSAFP levels, in the absence of other evidence of B19 infection, should not prompt Bl9-specific testing.

KEY woms-Human parvovirus B19, prenatal screening, maternal serum alpha-fetoprotein.

INTRODUCTION

Human parvovirus B 19 (B 19) was discovered serendipitously in 1975 (Cossart et al., 1975). Thereafter, various clinical entities-including erythema infectiosum, transient aplastic crisis, and chronic anaemia-have been attributed to infection with this virus.

In 1984, two reports appeared describing intra- uterine B 19 infection associated with unfavourable pregnancy outcomes (Knott et al., 1984; Brown et al., 1984). B19’s proclivity for and destruction of

Addressee for correspondence: David R. Johnson, MD, Michigan Department of Public Health, Bureau of Infectious Disease Control, 3500 N. LogadMLK Jr. Blvd., P.O. Box 30035, Lansing, MI 48909, U.S.A.

0 1994 by John Wiley & Sons, Ltd. CCC 0197-3851/94/060455-04

erythrocyte precursor cells may lead to severe fetal anaemia, congestive heart failure, hydrops, and fetal death (Young et al., 1984; Anand et al., 1987; Anderson et al., 1988).

Several cases of documented fetal B19 infection have been associated with high maternal serum alpha-fetoprotein (MSAFP) levels, prompting rec- ommendations that MSAFP be monitored to assess damage to the infected fetus (Carrington et al., 1987; American Academy of Pediatrics, 1990).

Reliable serological assays for human parvo- virus have been developed (Anderson et al., 1986), but are not yet widely available (Centers for Disease Control, 1989). Conversely, MSAFP sam- pling during the second trimester of pregnancy is a common practice and has been increasingly used to screen for certain fetal abnormalities,

456 D. R. JOHNSON ET A L

particularly neural tube defects. We hypothesized that an elevated screening MSAFP measure- ment might be an appropriate indication for B19- specific testing. We conducted a retrospective, matched, case-control study to evaluate this hypothesis.

MATERIALS AND METHODS

From 1 December 1986 to 31 January 1989, 21 392 women from throughout the state of Michigan had a serum sample drawn about the 16th week of pregnancy and submitted to the AFP Screening Program of Michigan State University (East Lansing, Michigan) for MSAFP testing. Each MSAFP result was converted into a multiple of the median (MOM) value for the appropriate gestational age following standard methods (Adams et al., 1984) and then adjusted for mater- nal weight. Sera remaining after the MSAFP test- ing were stored at - 20°C. Follow-up pregnancy outcome information was obtained on approxi- mately 40 per cent of the women, including nearly all those with elevated MSAFP levels.

Cases were defined as women who met the following criteria: MSAFP measurement greater than 3.0MOM; pregnancy which ended with an unplanned delivery of a stillborn fetus prior to 28 weeks’ gestation; and delivery of a stillborn with- out gross abnormalities.

Controls were those whose MSAFP levels were between 0.4 and 2.2 MOM, who had live births at >37 weeks’ gestation, and whose newborns had no gross abnormalities. Two controls were matched to each case on the following: collection of MSAFP sample within 3 months, residence in the same or contiguous county, race, previous delivery of liveborn infant(s), and maternal age within 5 years.

Serum samples from the case-patients and con- trols were tested for B19-speciiic IgM and IgG antibodies by enzyme-linked immunosorbent assay according to the procedure of Anderson et al. (1986). The laboratory technicians were blinded to the case or control status of the samples.

The matched odds ratio and its 95 per cent confidence intervals were calculated by the meth- ods of Robins et al. (1986). Formulae compiled by Schlesselman (1982) were used for study power computations, with the assumption that B19 IgM antibody would be found in 1 per cent of the control Donulation.

Table I-Comparison of women with elevated and normal maternal serum alpha-fetoprotein (MSAFP) levels

Elevated Normal MSAFP MSAFP (n= 173) ( ~ 1 5 283)

Mean age (years) 26.2 26.3 White race (YO) 74.6 83.5 Median gravidity 2 2

*Odds ratio=0.58, 95 per cent confidence interval 0.40-0.83.

RESULTS

Of the 16 368 women who underwent MSAFP testing during the 26-month study period and for whom basic demographic and pregnancy data were available, 173 (1.0 per cent) had a MSAFP level greater than 3.0 MOM and 15 283 (93.2 per cent) had a level above 0.4 but below 2.2 MOM. Those with elevated MSAFP levels had the same mean age and median gravidity, but were less likely to be white than those with normal MSAFP levels (Table I).

Thirty women met all the case definition criteria. Appropriately matched controls and serum samples were available for 22 of those 30. The final study sample of 66 women (22 cases and 44 controls) had a mean age of 27.2 years, a median gravidity of 2, and 55 per cent were white.

In none of the 66 serum samples was B19- specific IgM antibody detected. IgG to B19 was found in 33 (50 per cent; 95 per cent confidence interval 38-62 per cent) of the samples. There was no significant association found between case/ control status and the presence of IgG antibody; the matched odds ratio was 0.77, with a 95 per cent confidence interval of 0.28-2.1 1 (Table 11).

The estimates of power to detect a two-fold and a five-fold increased odds of finding B19-specific IgM among cases compared with controls were only 9 and 25 per cent, respectively, in this study.

DISCUSSION

The association between gestational infection with human parvovirus and fetal loss has been substantiated by a number of investigators ( h a n d et al., 1987; Anderson et al., 1988; Camngton et

I I al., 1987; Gray et aZ., 1986; Lefrere et aZ.; 1986;

SCREENING MATERNAL SERUM ALPHA-FETOPROTEIN LEVELS 457

Table 11-Data layout for study of possible association between casekontrol status and the presence of B19- specific IgG antibody

No. of B19 IgG+controls 2 1 0

Cases B19 IgG+ 2 6 2 B19 IgG - 3 7 2

Matched odds ratio=0.77, 95 per cent confidence interval 0.28-2.1 1.

Woernle et al., 1987; Kinney et al., 1988; Public Health Laboratory Service Working Party on Fifth Disease, 1990). That risk has been estimated at 9 per cent for serologically-proven maternal B 19 infection during pregnancy (Public Health Lab- oratory Service Working Party on Fifth Disease, 1990). Confirming or excluding the diagnosis of gestational B19 infection is useful; treatment modalities for the symptomatic fetus may appro- priately be considered (Schwarz et a/., 1988; Naides and Weiner, 1989; Soothill, 1990; Pryde et al., 1992) and more complete counselling regarding future pregnancies may be possible.

Perhaps the most straightforward way to deter- mine B19 susceptibility or infection status would be to screen all pregnant women for B19 IgG and IgM antibodies. The limited availability of parvo- virus antigens and reliable serological assays makes such an approach impractical at present. Consequently, we had hoped that elevated MSAFP levels would correlate with evidence of recent B19 infection.

Our negative findings in no way refute the association between gestational B 19 infection and fetal loss. Nor are they incompatible with the assertion that B 19-induced fetal symptoms lead to increased MSAFP levels. With this study sample, our power to detect this latter association was very low.

In terms of their past experience with human parvovirus, our 66 study subjects appear to be fairly representative of the general adult popula- tion; their 50 per cent sero-prevalence of B19- specific IgG antibody is very similar to that found in studies of other adults (Cohen and Buckley, 1988). These study subjects may also be represen- tative of the larger pregnant population in terms of recent B19 infection. If so, the fact that none had B19 IgM antibody would suggest that human parvovirus infection in pregnancy is neither com-

mon nor an important cause of increased MSAFP levels. Alternatively, it is possible that some of our case-patients were infected with B19 very early in gestation and lost their IgM antibody to it by the time of the second-trimester tests. In any case, an elevated MSAFP level on a routine prenatal screening test alone would not appear to be a good indication for performing B 19-specSc tests.

These data support current recommendations to restrict B 19-specific testing during pregnancy to those women who have a known or suspected exposure to B19, who themselves have symptoms suggestive of B19 infection, or whose pregnancies are affected by unexplained hydrops fetalis. Indi- cations for expanded, even routine human parvo- virus testing prior to or during pregnancy may be more appropriate when reliable serological assays become more widely available and when safe, effective methods of prevention (e.g., vaccine, specific immune globulin) are developed.

ACKNOWLEDGEMENTS

We thank Patricia Clark and Jack Holwerda of the Michigan Department of Public Health and G. William Gary of the Centers for Disease Control for their performance of the B19 serological test- ing. We also thank Henry King and David Loomis of Michigan State University for specimen and data management assistance.

REFERENCES Adams, M.J., Jr, Windham, G.C., James, L.M.,

Greenberg, F., Clayton-Hopkins, J.A., Reimer, C.B., Oakley, G.P. (1984). Clinical interpretation of mater- nal serum alpha-fetoprotein concentrations, Am. J. Obstet. Gynecol., 148, 241-254.

American Academy of Pediatrics, Committee on In- fectious Diseases (1990). Parvovirus, erythema in- fectiosum, and pregnancy, Pediatrics, 85, 13 1-133.

Anand, A., Gray, E.S., Brown, T., Clewley, J.P., Cohen, B.J. (1987). Human parvovirus infection in pregnancy and hydrops fetalis, N. Engl. J. Med., 316, 183-186.

Anderson, L.J., Tsou, C., Parker, R.A., Chorba, T.L., Wulff, H., Tattersall, P., Mortimer, P.P. (1986). Detection of antibodies and antigens of human par- vovirus B 19 by enzyme-linked immunosorbent assay, J. Clin. Microbiol., 24, 522-526.

Anderson, M.J., Khousam, M.N., Maxwell, D.J., Gould, S.J., Happerfield, L.C., Smith, W.J. (1988). Human parvovirus B19 and hydrops fetalis (Letter), Lancet, 1, 535.

Brown, T., Anand, A., Ritchie, L.D., Clewley, J.P., Reid, T.M.S. (1984). Intrauterine parvovirus infection

458 D. R. JOHNSON ET AL.

associated with hydrops fetalis (Letter), Lancet, 2,

Camngton, D., Gilmore, D.H., Whittle, M.J., Aitken, D., Gibson, A.A.M., Patrick, W.J.A., Brown, T., Caul, E.O., Field, A.M., Clewley, J.P., Cohen, B.J. (1987). Maternal serum alpha-fetoprotein-a marker of fetal aplastic crisis during intrauterine human parvovirus infection, Lancet, 1, 433-435.

Centers for Disease Control (1989). Risks associated with human parvovirus B19 infection, M.M. W.R 38,

Cohen, B.J., Buckley, M.M. (1988). The prevalence of antibody to human parvovirus B19 in England and Wales, J. Med. Microbiol., 25, 151-153.

Cossart, Y.E., Field, A.M., Cant, B., Widdows, D. (1975). Parvovirus-like particles in human sera, Lancet, 1, 72-73.

Gray, E.S., Anand, A., Brown, T. (1986). Parvovirus infections in pregnancy, Lancet, 1,208.

Kinney, J.S., Anderson, L.J., Farrar, J., Strikas, R.A., Kumar, M.L., Kliegman, R.M., Sever, J.L., Hurwitz, E.S., Sikes, R.K. (1988). Risk of adverse outcomes of pregnancy after human parvovirus B19 infection, J. Znfect. Dis., 157, 663467.

Knott, P.D., Welply, G.A.C., Anderson, M.J. (1984). Serologically proved intrauterine infection with parvovirus, Br. Med. J . , 289, 1660.

Lefrere, J.J., Dumez, Y., Courouce, A.-M., Deschene, G. (1986). Intrauterine infection with human parvo- virus, Lancet, 1, 449.

Naides, S.J., Weiner, C.P. (1989). Antenatal diagnosis and palliative treatment of non-immune hydrops

1033-1034.

81-88, 93-97.

fetalis secondary to fetal parvovirus B19 infection, Prenat. Diagn., 9, 105-1 14.

Pryde, P.G., Nugent, C.E., Pridjian, G., Ban, M., Faix, R.G. (1992). Spontaneous resolution of nonimmune hydrops fetalis secondary to human parvovirus B19 infection, Obstet. Gynecol., 79, 859-86 1.

Public Health Laboratory Service Working Party on Fifth Disease (1990). Prospective study of human parvovirus (B19) infection in pregnancy, Br. Med. J., 300, 1166-1 170.

Robins, J., Greenland, S., Breslow, N.E. (1986). A general estimator for the variance of the Mantel- Haenszel odds ratio, Am. J. Epidemiol., 124,719-723.

Schlesselman, J.J. (1982). Case-Control Studies, New York: Oxford University Press, 144-154.

Schwartz, T.F., Roggendorf, M., Hottentrager, B., Deinhardt, F., Enders, G., Gloning, K.P., Schramm, T., Hansmann, M. (1988). Human parvovirus B19 infection in pregnancy (Letter), Lancet, 2, 566-567.

Soothill, P. (1990). Intrauterine blood transfusion for non-immune hydrops fetalis due to parvovirus B19 infection (Letter), Lancet, 1, 121-122.

Woernle, C.H., Anderson, L.J., Tattersall, P., Davison, J.M. (1987). Human parvovirus infection during preg- nancy, J. Infect. Dis., 156, 17-20.

Young, N., Harrison, M., Moore, J., Mortimer, P., Humphries, R.K. (1984). Direct demonstration of the human parvovirus in erythroid progenitor cells infected in vitro, J. Clin. Invest., 74, 2024-2032.